*** nickr has joined #gnuenterprise
*** dcmwai has quit IRC
*** nickr_ has joined #gnuenterprise
*** nickr has quit IRC
*** reinhard has joined #gnuenterprise
*** dcmwai has joined #gnuenterprise
*** dcmwai has quit IRC
*** dcmwai has joined #gnuenterprise
*** johannesV has joined #gnuenterprise
*** dcmwai has quit IRC
*** dcmwai has joined #gnuenterprise
*** dcmwai has quit IRC
*** dcmwai has joined #gnuenterprise
*** dimas has joined #gnuenterprise
*** dcmwai has quit IRC
*** holycow has quit IRC
*** btami has joined #gnuenterprise
*** holycow has joined #gnuenterprise
*** SachaS has joined #gnuenterprise
<SachaS> good morning
<johannesV> good morning SachaS
*** btami has quit IRC
<SachaS> hallo johannesV
<SachaS> hey ... i just saw the commits, are those bound procedures working?
*** kilo_ has joined #gnuenterprise
*** btami has joined #gnuenterprise
<johannesV> yes, they're working
<johannesV> except the OnInit is missing
<johannesV> this is because this needs some more 'think-about' :))
<SachaS> is that true johannesV?
<johannesV> (conceptual stuff)
<SachaS> thats awesome
<SachaS> haha
<SachaS> :)
<SachaS> johannesV did u see my comment about timed procedures?
<SachaS> sort of time bound procedures
<SachaS> i saw there is a sort of python cron implementation
<johannesV> SachaS: no, should I? Where was it posted ?
<SachaS> how do you like that idea?
<SachaS> so we could have a gnue python cron job, which calls gnue-appserver procedures
<SachaS> like running some reports at night or whatever
<SachaS> let me get that url again
<johannesV> well, as i see things this isn't an issue of 'appserver' but
<johannesV> could be done with any system supporting cron-jobs
<johannesV> just create a 'client-app' (like appserver/src/testRPC.py) and call it via cron
<SachaS> true
<johannesV> this 'client-app' calls procedures from given classes
<SachaS> as long as you have the cron application on the host
<johannesV> and, running a report isn't a task from appserver but from report-server :)
<SachaS> :) psst :)
<SachaS> haha
<SachaS> or doing some analysis :)
<ajmitch> hi SachaS, johannesV
<SachaS> is that a report ? ;)
<SachaS> <SachaS> or doing some analysis :)
<SachaS> damn wrong key
<SachaS> http://sourceforge.net/projects/pycron/
<SachaS> hi ajmitch
<johannesV> SachaS: so this pycron-stuff runs on the client-side, right ?
<SachaS> might run on either side, actually ... was primarly thinking on the server side...
<SachaS> i dont have a use case but thought this might be another sort of bound procedure ...
<johannesV> well, ok, so this could provide a cron-like system to an os lacking such support
<SachaS> maybe, to get currency exchange rates into a gnue-appserver application. get the latest exchange rates, every day
<johannesV> but how does this affect appserver in the meaning of bound procedures
<johannesV> i mean where's the diff between a 'python script' and a 'pycron script' ?
<johannesV> both running 'somewhere' but connecting to the appserver
<SachaS> well, you could have a crontab file, like you have today. so you dont have to write such an app.
<SachaS> ok i understand what you mean..... i think
<SachaS> i could have a script, a gnue-appserver client, which gets the latest currency rates and then connects to the appserver and creates new CurrencyObjects on the appserver
<SachaS> i thought it could be nice, I have to admit I did not think much further than that.
<SachaS> johannesV: yesterday I went to the SAP NetWeaver WorldTour presentation.
<SachaS> i had to admit, they do cool stuff.
<johannesV> yeah, have read it in irc-log
<SachaS> no wonder, they about 2500 people working on NetWeaver
<johannesV> :)
<SachaS> so .... eventuelly there are cool things developing
<SachaS> ok.
<SachaS> basically
<SachaS> NetWeaver is all about integration
<SachaS> integrating different applications
<SachaS> integrating stuff
<ajmitch> 2500? wow
<ajmitch> that's even more than gnue has
<SachaS> 2500 yeah. amazing. :) 
<SachaS> anyway
<SachaS> the introduce a bus
<SachaS> the netweaver bus (not sure if the terminology is right)
<SachaS> anyway
<SachaS> there is like a bus
<SachaS> and applications connect to the bus
<SachaS> so application 1 just created a purchase order
<SachaS> so application 1 sends the information about the purchase order to the bus
<SachaS> any other application connected to the bus, can subscribe for information
<SachaS> so each application, which is subscribed to the purchase order information, gets that information.
<johannesV> ah, I see
<SachaS> so application 2 gets the information about the purchase order.
<SachaS> actually
<SachaS> this is not new
<SachaS> tibco has such a bus for quiet a while
<johannesV> well, i have to say this idea is quite good
<SachaS> tibco, how I understood even modified TCP/IP, the underlying transport protocol, to remove TCP/IP overhead! and to optimize TCP/IP for their bus. i thought thats crazy but would also make sense.
<SachaS> i like it, that you like it johannesV :)
<SachaS> ok.
<SachaS> so
<SachaS> on my diagramm. application 1 is a gnue-appserver
<johannesV> think of classes registering themselfs for notifications for changes of other classes ... :)
<johannesV> or events alike ...
<SachaS> application 2 is a ebXML Business Service Interface (what I developed as GPL at http://www.schlegel.li/ebXML/)
<SachaS> there is xmlblaster, which is such a bus. xmlblaster is open source, is implemented in java (does not really matter).
<SachaS> yes.
<SachaS> so, what we need
<SachaS> is a) a bus (maybe take XMLBlaster)
<SachaS> b) connectors
<SachaS> c) an interface for gnue-appserver
<SachaS> b) connectors: this is bus to gnue glue
<SachaS> lets say on the bus we have xml messages of the type UBL (Universal Business Language)
<SachaS> so the connector, if gnue-appserver generates a purchase order (in gnue purchase order structure), the connector would transform the gnue purchase order to a UBL purchase order.
<SachaS> the connector of the ebXML BSI application takes the UBL purchase order and transforms it to a (whatever) maybe keeps the UBL purchase order
<SachaS> c) the interfaces: the gnue-appserver must have an interface. so that if the connector has transformed the UBL pruchase order to a gnue purchase order, so the connector can send the gnue purchase order to the gnue system.
<SachaS> this could be a xml-rpc/rpc/soap/wsdl/gnue-messages interface
<johannesV> hmmm, the last transformation (ubl to gnue) would be easier than the other way round
<SachaS> for c: i could do the following: take the appserver testRPC.py sample and add rpc server to it. when the rpc server. the rpc server could have a procedure: incomingPurchaseOrder( gnue-purchase-order-type purchase-order). then the rpc server connects to the gnue-appserver, as it does now, and calls a "createPurchaseOrder" gnue-appserver procedure.
<johannesV> and where is the translation from ubl to gnue performed ? 
<SachaS> so, what we need is: to provide an interface to gnue-appserver business objects.
<johannesV> i think the testRPC -like app can do this transformation without using a lot of gnue-procedures
<SachaS> ubl2gnue would be in the xmlblaster-gnue connector
<johannesV> the other way round would be more complicated ... 
<SachaS> all the gnue-appserver interfaces do is to provide a rpc interface to the business objects
<SachaS> the xmlblaster-gnue connector makes sure it calls the right rpc gnue-appserver procedures
<johannesV> for a first try you could use the OnValidate-Event of gnue-purchase-order business object, since this get's called before a commit
<johannesV> but, this has no means of transactions ... 
<johannesV> so if a single  validation in a commit fails, there would have been a few other instances before calling a transformation-service
<SachaS> johannesV, let me start again.
<SachaS> we have the appserver
<SachaS> we have a business class called PurchaseOrder
<SachaS> so far
<johannesV> ok
<SachaS> we had gnue-forms and a user could fill out the information to create a purchase order
<SachaS> the gnue-form calls the createPurchaseOrder, or the PurchaseOrder constructor, to create a purchase order.
<SachaS> then the gnue-form make a commit
<SachaS> before the commit, the OnValidate procedure kicks in
<SachaS> and validates
<SachaS> if ok, the commit is executed
<SachaS> if it fails, probably an exception is thrown back to the gnue-form
<SachaS> johannesV, is that how it works now?
<johannesV> right, but validate is executed for all instances changed
<SachaS> so here we have the create purchase order transaction
<SachaS> hmmmm
<SachaS> i am currently only interested in the "to be created" purchase order
<SachaS> not if the user had changed objects in his gnue-form ...
<SachaS> it might not matter, actually
<SachaS> or do i miss somethign?
<SachaS> oh i see.
<johannesV> no, just continue; so the validate-event of PurchaseOrder calls the transformation to ubl
<SachaS> i think that does not matter.
<SachaS> no transformation yet
<SachaS> this is how it works today
<SachaS> ok?
<SachaS> no transformation necessary here
<SachaS> the gnue-form calls straight the gnue-appserver purchase order creat procedure
<SachaS> we are all in gnue land here
<SachaS> ok?
<SachaS> all gnue land here
<SachaS> ok johannesV?
<SachaS> so far we had only gnue.
* johannesV phone
<SachaS> bah
<SachaS> :)
<johannesV> ok, back
<SachaS> www.xmlblaster.org is the message oriented middleware software. has a xmlblaster licence, which is based on LGPL licence ...
<SachaS> ok
<SachaS> so far we have been in gnue land
<SachaS> the next step
<SachaS> would be to provide a non gnue-forms interface to the appserver
<SachaS> lets say a rpc interface
<SachaS> through this interface
<SachaS> a normal rpc client can create a purchase order
<SachaS> lets call this interface gnue-appserver-interface
<johannesV> acutally there is an rpc interface to the appserver
<SachaS> i could imagine, that in the .gsd file, one could specify, whether a business object shall be available through the non gnue-form interface, anyway.
<SachaS> what rpc interface is that?
<johannesV> gnue-forms talks to gnue-appserver via xmlrpc
<johannesV> for all clients, there's no other way to talk to appserver but rpc
<SachaS> ok i dont know enough how, under the hood, gnue-forms talks to gnue-appserver
<SachaS> what i need is a situation, where i can write a rpc client, which does not have to import any gnue.something imports
<SachaS> i need a non gnue rpc client
<SachaS> i would need something
<SachaS> where i can have my ruby rpc client
<SachaS> which calls a rpc procedure of the gnue-appserver
<SachaS> hmmmm
<johannesV> ok, i see, but this shouldn't be a problem (since the appserver-api is documented)
<SachaS> lets say no security, no login,
<johannesV> we would have to talk to reinhard for such a rpc-specific thing
<SachaS> so for example
<SachaS> ok
<SachaS> do you have some more time?
<johannesV> of course
<SachaS> ok
<SachaS> lets say, I want the PurchaseOrder business object to be accessible through the non-gnue-rpc interface"
<SachaS> i set a flag in the .gsd file
<SachaS> somehow, the gnue-appserver gets the information to provide a public interface of the PurchaseOrder business object
<SachaS> ok. the gnue-appserver, then creates a rpc procedure called: boolen purchaseOrder( some, arguments )
<SachaS> so with my normal rpc client
<SachaS> i can call this purchaseOrder rpc procedure
<johannesV> i would do that in a more generic way (I'm not sure but i think this is the way it was designed)
<johannesV> like request the appserver to create an instance of class 'PurchaseOrder'
<SachaS> the purchaseOrder rpc procedure then creates a PurchaseOrder business object.
<SachaS> exactly
<SachaS> my point is
<johannesV> and doing this with 'PurchaseOrder' as argument
<johannesV> so you're able to create new instances of 'all' classes
<SachaS> the client must only speak rpc
<SachaS> the client must not import any gnue stuff
<johannesV> ok, but that doesn't matter
<SachaS> ok
<SachaS> than thats good
<SachaS> :)
<johannesV> right, there's no need to, because "create a new instance of ..." is already a 'rpc-function'
<SachaS> ok
<SachaS> next step is the connector
<SachaS> ok.
<SachaS> maybe I want to have purchaseOrderCreate( foo, bar) instead of create_object( purchaseOrder )..... but maybe this is no differece at all...
<SachaS> ok
<SachaS> next step is the connector
<SachaS> lets say the connector gets some strange ;) UBL purchase order message
<SachaS> the connector must know how to read the UBL purchase order message
<SachaS> in our example
<SachaS> the connector must know
<SachaS> that when it receives a purchase order message (UBL or EDI or whatever), it must call the "create a new instance of Purchase Order" rpc procedure at the gnue-appserver.
<SachaS> the connector must also know
<SachaS> how to parse the incoming purchase order message
<SachaS> the connector parses (or maps) the information of the UBL purchase order message, and can create the right arguments for the  "create a new instance of Purchase Order" rpc procedure
<SachaS> :)
<johannesV> ok, or whatever mechanism is provided to create such an instance
<SachaS> exactly
<johannesV> but this is only the way from outside into the appserver; what about the way out from appserver
<SachaS> ok.
<SachaS> i am not sure
<SachaS> but i think
<SachaS> it was not the idea of gnue-reports
<SachaS> but currently
<SachaS> gnue-reports can send an email for example
<SachaS> so it could also call a rpc procedure
<SachaS> but basically
<SachaS> lets take gnue-reports
<SachaS> lets say, the gnue-appserver (based on a bound procedure of the inventory module) creates a purchase order
<SachaS> so a business object procedure will create a purchase order message ( lets say a gnue-purchase-order )
<SachaS> i had the evil thought of a procuedre could actually make a rpc call ;)
<SachaS> anyway
<SachaS> either a procedure or gnue-reports
<SachaS> calls the outgoingPurchaseOrder rpc procedure of the connector
<SachaS> then
<SachaS> the connector makes the transformation
<SachaS> the information is somewhere (maybe passed in the rpc procedure as argument)
<SachaS> that the end system wants a UBL message
<SachaS> so the connector (or the transformation manager of the connector)
<SachaS> parses the gnue purchase order
<SachaS> and creates (maps) a UBL purchase order
<SachaS> next
<SachaS> the connector literally throws the UBL purchase order message on the bus
<SachaS> and any application, which is subscribed to a purchase order, receives the UBL purchase order
<johannesV> but wouldn't this create a recursion ?
<SachaS> sorry. the gnue conector (allways) transforms to UBL (or whatver the one type is chosen for the bus)
* holycow waves
<holycow> interesting conversation
<johannesV> at the beginning of your example, appserver was listening to such messages
<SachaS> hold on
<SachaS> yes
<SachaS> you get the point
*** holycow has quit IRC
<SachaS> but hold on
<SachaS> yeah somehow we would have to make sure
<SachaS> this time the appserver connector throws the message on the bus
<SachaS> so it should not receive this message
<SachaS> but yeah
<SachaS> have not thought about it
<SachaS> a bus might handle this
<SachaS> or we might be talking about some different types of PurchaseOrders
<SachaS> not sure
<SachaS> but
<SachaS> my ebXML Business Service Interface receives the PurchaesOrder, which was created by the gnue-appsever
<SachaS> and the ebXML Business Service Interface then sends the purchase order to the other party (the party the purchase order is for).
<SachaS> i think you got the point.
<SachaS> and i think this is pretty neet
<SachaS> pretty nice
<SachaS> for example
<SachaS> a statistics tool can also subscribe to purchase orders
<johannesV> hmmm, as i see things this concept needs some more tlc
<johannesV> but it's interessting
<SachaS> so not only the ebXML BSI system gets the purchase order but also the statistical tools
<SachaS> i am not interested in what the statistical tool does with the purchase order
<SachaS> johannesV
<SachaS> what do think when I pass by this afternoon?
<SachaS> i have a car
<SachaS> would take me about 40 minutes
<SachaS> if you are living close to reinhard
<SachaS> :)
<SachaS> must be the phone
<johannesV> SachaS: actually I'm living close to reinhard, but he's sick today and so I've got the hotline-services today
<johannesV> right
<SachaS> i dont mind if you have the hotline service
<SachaS> :)
<johannesV> btw. there's another bunch of work to be done too
<SachaS> maybe after 6pm
<SachaS> if you are interested
<SachaS> anytime
<SachaS> i have to run for lunch
<johannesV> Sorry, but after 5pm there's 'family time' -- maybe another time ...
<SachaS> i will see you after lunch, I hope
<SachaS> what you have a life?
<SachaS> :)
<johannesV> :)
<SachaS> oh i see
<SachaS> well see you after lunch
*** ajmitch__ has joined #gnuenterprise
<kilo_> reinhard, johannesV: does gacvs --selftest run OK at you?
*** reinhard_ has joined #gnuenterprise
*** reinhard has quit IRC
<johannesV> kilo_: yes it's ok
<kilo_> thx. postgreSQL isnt it?
<johannesV> kilo_: yes, PostgreSQL 7.4.2
<johannesV> what are you using ?
<kilo_> interbase and we have a major problem here
<kilo_> cannot start appserver at all
<SachaS> whats the url for the irc-log ?
<kilo_> www.gnuenterprise.org/irc-logs/
<SachaS> thanks kilo_
* SachaS goes and updates babyerp
<SachaS> can someone please kick svn.
<btami> we have to wait for jamest to kill apache-ssl or likes
<SachaS> ok
<SachaS> johannesV: question concering procedure definition
<SachaS> if i choose a boolean datatype
<SachaS> what values do i need for scalable and length?
<kilo_> dont bother him, he is working very hard! 8-)))
<SachaS> what values shall i put in there
<SachaS> :) true
<SachaS> kilo_ I want to update babyerp and then try calculated fields and bound procedures
<kilo_> oh, you lucky. i want to start appserver...
<SachaS> :)
<SachaS> actually what I experience when editing the .gsd file by hand ... its a mess. its difficult to understand. maybe i should cut it and put each module into one .gsd file ...
*** dsmith has joined #gnuenterprise
<johannesV> SachaS: nothing, boolean don't want to have sclae and length
<SachaS> so           <value field="gnue_length"></value>
<SachaS>           <value field="gnue_nullable">True</value>
<SachaS>           <value field="gnue_scale"></value>
<SachaS>           <value field="gnue_type">boolean</value>
<SachaS> is ok?
<SachaS> this is from a procedure.
<SachaS> or even remove the <value> element?
*** btami has quit IRC
*** reinhard_ has quit IRC
*** kilo_ has quit IRC
*** jamest has joined #gnuenterprise
<johannesV> SachaS: back now ... just installed a firebird-server on my debian ...
<johannesV> compared to postgresql dbmanagement seems to be much more complicated with firebird
<johannesV> SachaS: you could leave the <value>-tags away if they're empty
*** reinhard has joined #gnuenterprise
<johannesV> this usually means they're null
<johannesV> howdy reinhard
<SachaS> ok.
<SachaS> thanks.
<reinhard> hi all
<SachaS> hi reinhard
<johannesV> reinhard: just installed firebird on boston :)
<johannesV> kilo can't start appserver anymore ...
<SachaS> i used to return lists all the time in babyerp ... have to change it. 
<SachaS> i feel like throwing babyerp away and starting from scratch
<SachaS> but thats too much work
<SachaS> bah its hard to maintain babyerp ;)
<SachaS> the babyerp management sucks
<SachaS> the babyerp project management sucks
<SachaS> having a list as return value would be great :)
*** johannesV_ has joined #gnuenterprise
*** dneighbo_ has joined #gnuenterprise
<SachaS> when the gnue packages get started again (with appserver)
<SachaS> coding policies must be in place.
<SachaS> jamest: can you kick svn for anonymous update?
<SachaS> please
<SachaS> sacha@gnaraloo:~/gnuenterprise$ svn up gnue
<SachaS> svn: Berkeley DB error while opening 'nodes' table for filesystem /var/svn/gnue/db:
<SachaS> Cannot allocate memory
<andrewm> nasty
<andrewm> ash has lots of memory free
*** johannesV has quit IRC
<jamest> sigh
<jamest> it's the nightly svn dump i think
<jamest> seems to hang
<jamest> sigh
<jamest> today that's not the case
<dsmith> jamest: WHat user does svn processes run as?
<dsmith> jamest: Maybe there is a resource limit that needs to be adjusted?
<jamest> runs as svnserve
<jamest> unless it's svn+ssh
<jamest> then it runs as that user
<jamest> so the wedged anon svn shouldn't effect ssh checkouts which it seems to
<jamest> it's up again btw
<SachaS> ta
*** dneighbo has quit IRC
*** dneighbo has joined #gnuenterprise
<dsmith> jamest: There are probably lock files that keep it from being accessed.
<SachaS> the sample in gnue-appserver/samples has a problem
<SachaS> i can create the db
<SachaS> appserver starts
<SachaS> then i start sample.gfd form
<SachaS> and run a query
<SachaS> i get an exception.
<SachaS> i get a fresh copy of gnue-appserver first...
<SachaS> might have been a mistake from my part ... will try again
<johannesV_> SachaS: will test it 
<SachaS> ok. working
<johannesV_> SachaS: which db are you using ?
<SachaS> i did not create new database sql file from .gsd file ...
<SachaS> now is working
<johannesV_> ah, ok 
<SachaS> johannesV let me know when you have time, maybe when reinhard is no longer hoased we can have a chat.
<johannesV_> SachaS: yeah, maybe tomorrow
<SachaS> i am not here tomorrow or friday.
<reinhard> vcs
<reinhard> grrr
<SachaS> i can make it saturday, sunday or next week
<SachaS> hi hoarsed reinhard
<reinhard> i will be not here saturday or sunday
<SachaS> ;) wrong window
<reinhard> if i'm still ill, i will be in bed
<SachaS> so shall we meet next week somewhen?
<reinhard> if i'm ok by then, i will make a trip with the family
<reinhard> better next week, yes
<SachaS> ok
<johannesV_> ok
<SachaS> shall we all meet?
<SachaS> shall we meet at reinhards place?
<SachaS> during the day, maybe
<reinhard> i would prefer meeting in the evening
<SachaS> johannesV has a life
<reinhard> if it is ok for johannesV_
<reinhard> :)
<SachaS> so might be a bit difficult for him ;)
<SachaS> its fine with me
<reinhard> i think we should decide next week
<SachaS> ok
<reinhard> monday is holiday
<reinhard> so let't talk again tuesday, ok?
<SachaS> ok. so no monday
<SachaS> ok
<SachaS> yes.
<johannesV_> ok, guys, so I'm away for about an hour ...
<johannesV_> have fun :)
<SachaS> ok
<SachaS> is the classrepository updated for calculated properties?
<reinhard> i think yes
<reinhard> calculated properties are just procedures with the name starting with "get"
<SachaS> so if i have a procedure: getMyAge
<reinhard> yes
<SachaS> or getmyAge
<SachaS> i can have:
<SachaS> my_business_object.myAge
<SachaS> ?
<reinhard> exactly
<SachaS> ok. so i best have a getmyAge procedure
<SachaS> to get the property name lower case
<SachaS> ok. will try
*** dsmith has quit IRC
*** dsmith` has joined #gnuenterprise
<SachaS> when using gnue designer and connecting to appserver, there is no introspection anymore
<SachaS> let me check again
<SachaS> yep no introspection anymore
<SachaS> the "Business Object Class" tree stays empty
<SachaS> actually i get a traceback
<SachaS> DB000:   File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/com mon/rpc/drivers/Base.py", line 232, in __call__
<SachaS> DB000:     return self._adapter._runMethod (self._methodname, *args, **params)
<SachaS> DB000:   File "/home/sacha/gnuenterprise/gnue/gnue-common/.cvsdevelbase/gnue/com mon/rpc/drivers/xmlrpc/pw_xmlrpc/ClientAdapter.py", line 96, in _runMethod
<SachaS> DB000:     raise client.DistantError, e.faultString
<SachaS> DB000: gnue.common.rpc.client.DistantError: 'dict' object has no attribute '_chi ldren'
<SachaS> DB000:
<SachaS> maybe i do a fresh svn checkout ...
<SachaS> first
<SachaS> yes. introspection of gnue-appserver is borked
<dneighbo> does anyone know if you can set multi fields for a m/d setup?
<dneighbo> i.e. if i have
<dneighbo> fiscal year and accountcode in a ledger table
<dneighbo> and i have fiscal year and accountcode in foo table
<dneighbo> and i want to make foo a "detail" datasource to ledger
<dneighbo> can i say
<jcater> iirc, yes
<jcater> comma-separated
*** andrewm has quit IRC
<dneighbo> master field = accountcode,fiscalyear
<jamest> that's what i was thinking too
<dneighbo> SWEET
<jcater> masterlink="key1,key2"
<jcater> detaillink="key1,key2"
<jamest> i'm 99% sure it works in reports anyway
* dneighbo does the super uber ass happy dance
<jcater> it did work
<jcater> for what that means :)
<jcater> i.e., not much
* dneighbo was having night terrors last night that i was going to have to add columns to my tables
<jamest>  /msg jcater boy, he has a pretty high opinion of his ass doesn't he?
<dneighbo> jamest you follow postgres fairly closely
<dneighbo> do you know what version the expect views to be updateable?
<jamest> um
<jamest> i'm not even sure that's in the works
<jamest> but a view is IIRC just a rule
<jamest> so you can IIRC create updateable views by adding rules to them
<dneighbo> ok the multi key m/d appears to at least not croak syntax wise. :)
<dneighbo> jamest: i will have our dba research
<jcater> I thought non-joined views were already updatable in postgres
<dneighbo> at this point its not critical, we just like the concept of views
<jamest> jcater: if so then that's a recent addition
<dneighbo> we understand joined views can not be updated but were disappointed non joined wouldnt
<dneighbo> jcater: all documentations says no
<jcater> I've never used it
<jcater> so may be wrong
<dneighbo> and 7.4 does seem to allow
<jcater> that's one of those things I'd never do
<jcater> so wouldn't have tried
* dneighbo always thought it was possible, but never tried
<jamest> Currently, views are read only: the system will not allow an insert, update, or delete on a view. You can get the effect of an updatable view by creating rules that rewrite inserts, etc. on the view into appropriate actions on other tables. For more information see CREATE RULE. 
<dneighbo> jcater: why would you never do
<jcater> just defeats my own internal view of what a view is
<jcater> err, no pun intended
<dneighbo> gotcha
<jamest> i'm w/ jcater on that one
<dneighbo> in the past i pretty much always considered views read only
<dneighbo> i think some of the reason is concerns about gnue limitations that had us looking at views that needed updating
<dneighbo> however, i think we have fixed some of those problems so wont be necessary
<dneighbo> jcater: quick bug report (designer)
<dneighbo> there appears no way to delete a datasource or block in designer
<jcater> won't rules do what you need?
<dneighbo> jcater: yes
<dneighbo> they will
<dneighbo> but at this point i dont htink we will even need them
<jcater> I'm not happy w/designer
<jcater> I have plans
<dneighbo> my question was more general in curiousity (as i agree that updateable views should try to be avoided)
<jcater> but no time atm
<dneighbo> what kind of plans
* dneighbo is fearful now
<dneighbo> as i actually have 3 staff learning designer and getting comfortable :)
<jcater> don't worry
<jcater> polish
<jcater> internal stuff
<jcater> it's just not "cohesive"
<dneighbo> despite its perks if it changes too radically i might have a jamest like user uprising on my hands
<jcater> one of the main changes I want to do
<jcater> which I've flipped-flopped on since the beginning
<dneighbo> this transition to gnue has been painful for them
<dneighbo> not because of anything gnue related just new language, new to event driven (visual) programming and new to sql
<jcater> I've tried to use gnue-form's wx driver to draw the widgets in the layout editor pane
<jcater> that's just not working out
<jcater> having native widgets in that pane
<jcater> vs virtual ones that I draw
<jcater> is no longer worth the headache for me
<dneighbo> :)
<jcater> there's so many features I want in that pane
<jcater> that I can't do w/the currnet method
<dneighbo> cool
<jcater> like clicking edge of a field
<jcater> and dragging it to make it bigger
<jcater> or sane keyboard shortcuts
<dneighbo> well i suspect yo will get lots of "bugs" (ie feature enhancements) from us on designer
<dneighbo> we keep breakign forms and then solving the problem.. so no good bug reports comign from there
<jcater> plus it kills me that I *have* to use the mouse in designer
<jcater> I'm all for gui interfaces
<dneighbo> though we hit an odd on ethe other day
<jcater> but once I use a tool alot
<jcater> I need keyboard shortcuts
<dneighbo> jcater: agree 100%
<jcater> which I've tried to do
<dneighbo> we had a timestamp field
<dneighbo> in forms and in postgres
<jcater> but the current model of the widgets makes that hard
<dneighbo> we change a text field on the form
<dneighbo> and hit "save" (commit"
<dneighbo> it has no errors and is happy
<dneighbo> we shut the form, reopen (or requery)
<dneighbo> and the change didnt save
<dneighbo> so we debug and check the query built in forms
<dneighbo> we paste in psql and run
<dneighbo> it runs WITHOUT error
<dneighbo> but retures 0 ROWS UPDATED
<dneighbo> wtf?
<jcater> jamest broke it
<dneighbo> come to find out timestamps in our database have seconds
<dneighbo> but forms doesnt
<dneighbo> so it didnt "error" but didnt "post"
<jcater> ah
* dneighbo thinks we need to somehow notify users that nothing updated
<jcater> I think that is an issue w/one of the dbdrivers
<jcater> losing seconds
<dneighbo> psycopg is what we are using
<jcater> I'd try pypgsql and see if that happens w/it too
<jcater> is this woody?
<dneighbo> im convinced this is not a "gnue" problem, BUT i think we need to display an error on this condition even though the database doesnt
<dneighbo> nope unstable
<dneighbo> as if i hit save
<dneighbo> and get 0 rows updated
<dneighbo> thats a problem :)
<dneighbo> as techincally teh only times forms allows me to "save" is if it thinks i changed a value 
<dneighbo> so there should be no reason i get a zero update
<dneighbo> if that makes sense
<dneighbo> we didnt need the time stamp on the form (We just had it there for some testing early on)
<dneighbo> so we removed it and references to it and we are good to go now
<dneighbo> but wanted to bring it up for thought
<SachaS> appserver calculated fields working for me.
<SachaS> oh dneighbo: the freshmeat entries to gnue-<xyz> have wrong urls
<SachaS> someone pointed out the other day
<dneighbo> sigh...
<dneighbo> we need bug tracking
* dneighbo will try to get dcl up during next two weeks
<SachaS> bug tracking would be good yes. specially if you have now real users
*** SachaS has quit IRC
<jcater> dneighbo: which windows svn client are you using?
<jamest> that's a db driver issue w/ timestamps and gnue/forms doesn't know it's happening
<jamest> as some of the postgresql drivers strip arbitrary amounts of resolution from the time
<jamest> so the db stores 10:30:02.12312 and the driver gives forms 10:30:02
<jamest> now forms says time to update and does a per field where clause on the update
<jamest> update foo set bar = 1 where time = "10:30:02"
<jamest> no errors will occure
<jamest> occur
<jamest> maybe we should check to see if the post altered as many rows as expected
<jamest> but I can't recall if python dbapi can give us such info
<dneighbo> jcater: tortoise and we dont like it
<dneighbo> how can i set the value of an entry
<dneighbo> block.entry.value = foo
<dneighbo> or 
<dneighbo> block.entry.set(foo)
* dneighbo knows it has changed some from last time i did it
<dneighbo> and dont readily see an form...
<dneighbo> looks like block.entry.set(foo) based on intro.gfd
<jcater> .set(foo) I know works
<dneighbo> grr parameters make no sense to me
*** btami has joined #gnuenterprise
<jcater> well, try to think of them like ducks
<reinhard> ducks?
<reinhard> you mean goats?
<jamest> shouldn't leave out the sheep either
<dneighbo> i got it figured out
<dneighbo> i was misreading the sample
<dneighbo> it was overly clever after breaking it down works just fine
<dneighbo> i looked at the api manual
<dneighbo> i need to pass a param and query off it
<dneighbo> do you know if
<dneighbo> blkFoo.initQuery
<dneighbo> blkFoo.Entry = getParameter("myParam")
<dneighbo> blkFoo.processQuery
<dneighbo> is a sane way to do such a thing?
*** johannesV_ has quit IRC
<jamest> shouldn't that bet a blkFoo.Entry.set(getParamter("myParam"))
<jamest> as the conversion via = is sometimes wack
<dneighbo> jamest: probably.. :) i was asking
<dneighbo> though i tried .set() and i think it failed
<jamest> i prefer the = myself for reading
<jamest> but i use set()
<dneighbo> fsck my mouse disappeared on this machien
<dneighbo> anyoen know of a way to get mouse to work again?
*** wendall911 has joined #gnuenterprise
<dsmith> dneighbo: 2.6.6 ?
<dsmith> dneighbo: No
<dneighbo> something like that
<dsmith> dneighbo: Mine went away whe I tried it the other day. I haven't spent any time tracking it down.
<dneighbo> im on 2.6.4-1-686
<dsmith> 2.6.6 just arrived in Sid the other day.  That's what I tried.  X *looked* like it found the mouse, but it wasn't responsive.
*** SachaS has joined #gnuenterprise
*** dneighbo has quit IRC
*** reinhard_ has joined #gnuenterprise
*** reinhard has quit IRC
*** reinhard_ has quit IRC
*** bluesbaron has joined #gnuenterprise
<jamest> dsmith: i had that issue
<jamest> seems that I had to load both the mouse kernel module and the ps2 mouse kernel module
<jamest> then X was happy again
<dsmith> jamest: THanks!
*** REAL_DEATH has joined #gnuenterprise
<REAL_DEATH> 'all americans and israels death'
<REAL_DEATH> 'all americans and israels death'
<nickr> lolz
<dsmith> Ops?
*** ChanServ sets mode: +o jcater
*** ChanServ sets mode: +o jamest
*** jcater sets mode: +b *!*@*.adsl.ttnet.net.tr
<jamest> bye
*** jcater sets mode: +b *!*@*.adsl.ttnet.net.tr
*** REAL_DEATH was kicked by jcater (jcater)
<jamest> wow
<dsmith> jamest: mousedev.ko and psmouse.ko ?
<jamest> i think so
<dsmith> Cool. I try it soon.
*** jcater sets mode: -o jcater
*** jamest sets mode: -o jamest
*** psu has joined #gnuenterprise
<jamest> hey psu
<psu> hi
<SachaS> hi psu
<bluesbaron> You guys really need to slow down. I have a hard time keeping up with the conversation ;) 
<bluesbaron> Anyway...hi all 
<bluesbaron> First time here, trying to get involved 
*** sjc has joined #gnuenterprise
<psu> bluesbaron : heh. of course, when I was doing the GNUe Traffic summaries, everyone was much more talkative
<psu> Cause or effect?
<bluesbaron> Won't burn myself on that on yet. Hey, I can't get to the roadmaps. Any particular reason? 
<psu> they used to be at http://www.gnuenterprise.org/docs/roadmaps.html on the old site
<bluesbaron> That's a big 404 
<SachaS> well
<SachaS> go to www.gnuenterprise.org
<SachaS> on the left you have one link called: Developer Tools
<SachaS> go to the developer tools
<SachaS> ooops
<SachaS> ;)
<SachaS> do:
<SachaS> svn co svn://svn.gnuenterprise.org/var/svn/gnue/trunk gnue
<SachaS> ;)
<SachaS> and you get all files, including the roadmaps
<SachaS> :)
<dsmith> When svn is working, of course.
<SachaS> yes of course
<SachaS> i thought they were online at one point in time
<jamest> they were
<jamest> i thought they still were
<SachaS> i cant find them
<jamest> http://www.gnuenterprise.org/tools/forms/
<jamest> on each tools page there is a roadmap link on the left
<SachaS> outch
<SachaS> yep
<psu> <homer>doh!</homer>
<jamest> bluesbaron: i think we're all pretty busy at the moment so not chating it up much
<jamest> both on the ML and in IRC
<SachaS> why does forms have so many todos and in progress and so few finished?
<SachaS> just kidding
* jamest slaps SachaS about with a goat
<jamest> i blame the babyerp team
<jamest> or derek
<bluesbaron> I think I adapted to the pace quite well ;) 
<bluesbaron> Thanks Sacha and James, I'll check it out 
*** psu has left #gnuenterprise
*** kilo_ has joined #gnuenterprise
<SachaS> where is johannesV? he just committed a bug fix
<SachaS> thanks johannesV
<SachaS> can someone kick svn, please?
<btami> jcater: can you explain your "focus order" plans in designer ?
<btami> i need them, and tried to finish some of code you started
<btami> just want to get some guidelines
<btami> is you don't mind...
<jcater> um
<jcater> what I want
<jcater> isn't easily doable w/the current way layout editor works
<jcater> which is why it got put on back burner
<jcater> I plan to have a 
<jcater> (1)
<jcater> (2)
<jcater> etc 
<jcater> (numbers in colored circles)
<jcater> appear on top of each widget
<jcater> and you can change this order by clicking on the widgets
*** dneighbo has joined #gnuenterprise
<jcater> when in "focus order mode"
<jcater> but wx won't let me easily do those colored circle things
<btami> ah
<btami> first i just need "Reset Focus Order by Position" to work
<btami> is 123 button in LayoutEditor's toolbar for switching "focus order" mode ?
<jcater> yes
<jcater> ah
<jcater> yeah, not sure when "Reset Focus Order by Position" stopped working
<jcater> it worked at one point
<jcater> then derek broke it
<btami> :)
<kilo_> these sentences dhould be written in big size fonts on GNUE website frontpage 8-)))
*** dsmith has quit IRC
*** btami has quit IRC
*** dsmith has joined #gnuenterprise
*** kilo_ has quit IRC
<SachaS> can someone please kick svn for anynmous update
<SachaS> jcater or jamest, can you kick svn please
<SachaS> appserver introspection is working again
*** sjc has left #gnuenterprise
<dsmith> jamest: Thanks again
<jamest> jcater: can you kick svn for me, i have to run
<jamest> and not sure if/when i'll make it back tonight
<dsmith> Linux TAG00604 2.6.6-1-686 #1 Wed May 12 14:57:57 EST 2004 i686 GNU/Linux
*** jamest has quit IRC
<dsmith> Yayy
<SachaS> dsmith, are you going to Linux Tag 2004?
<dsmith> SachaS: heh
<dsmith> Stoke: Silly name I have to have for my machine.
<dsmith> Time for home
*** dsmith has quit IRC
*** SachaS has quit IRC
*** bluesbaron has left #gnuenterprise
*** SachaS has joined #gnuenterprise
*** SachaS_ has joined #gnuenterprise
*** SachaS has quit IRC
<dneighbo> tortoise sux ass
<dneighbo> er windows sux ass
<nickr> ...
<jcater> SachaS_: should be working now
*** jcater has quit IRC
<SachaS_> ok. patched the file from the commit email ;)
<SachaS_> thanks
*** rsb_ has quit IRC
* SachaS_ is having issues with babyerp
<SachaS_> so i will drop babyerp, once again, for the moment
<SachaS_> and look at xmlblaster and ebXML business service interface.
<SachaS_> so i can have a message oriented middleware working with my ebxml bsi
*** jcater has joined #gnuenterprise
*** bluesbaron_ has joined #gnuenterprise
*** Morphous has quit IRC
*** Amorphous has joined #gnuenterprise
*** bluesbaron_ has left #gnuenterprise
<SachaS_> hmmm xmlbluster has a python library for python clients (gnue-appserver or gnue-appserver-interface) but no ruby library.
<SachaS_> so to get my ebxml stuff running with xmlblaster i might have to code the library in ruby...
<SachaS_> writing that library defenitely takes more than a blink of the eye
<SachaS_> maybe not that bad
*** mixi^ has joined #gnuenterprise
*** wendall911 has quit IRC
*** mixi has quit IRC
*** jamest has joined #gnuenterprise
*** jamest has quit IRC
*** Morphous has joined #gnuenterprise
*** Amorphous has quit IRC
*** jcater has quit IRC